A study of H I-selected galaxies in the Hercules cluster

A&A 406, 453–469 (2003)DOI: 10.1051/0004-6361:20030626c© ESO 2003

Astronomy&

Astrophysics

A study of H I-selected galaxies in the Hercules cluster�

J. Iglesias-Paramo1, W. van Driel2, P.-A. Duc3, P. Papaderos4, J. M. Vılchez5, V. Cayatte2, C. Balkowski2,K. O’Neil6, J. Dickey7, H. Hernandez6, and T. X. Thuan8

1 Laboratoire d’Astrophysique de Marseille, Traverse du Siphon – Les Trois Lucs, 13376 Marseille, Francee-mail: [email protected]

2 Observatoire de Paris, GEPI, CNRS-UMR 8111 and Universite Paris 7, 92195 Meudon Cedex, Francee-mail: [email protected]; [email protected]; [email protected]

3 CNRS URA 2052 and CEA, DSM, DAPNIA, Service d’Astrophysique, 91191 Gif-sur-Yvette Cedex, Francee-mail: [email protected]

4 Universitats-Sternwarte, Geismarlandstrasse 11, 37083 Gottingen, Germanye-mail: [email protected]

5 Instituto de Astrofısica de Andalucıa (CSIC), Granada, Spaine-mail: [email protected]

6 Arecibo Observatory, HC3 Box 53995, Arecibo, Puerto Rico 00612, USAe-mail: [email protected], [email protected]

7 Department of Astronomy, University of Minnesota, 116 Church Street SE, Minneapolis, MN 55455, USAe-mail: [email protected]

8 Astronomy Department, University of Virginia, Charlottesville, VA 22903, USAe-mail: [email protected]

Received 26 April 2002 / Accepted 9 April 2003

Abstract. The present study focuses on a sample of 22 galaxies detected in the blind VLA H survey of the Hercules clusterby Dickey (1997), 18 of which were selected on an H line width smaller than 270 km s−1 and 4 others with only tentativeoptical counterparts in the Palomar Sky Survey. Sensitive single-dish H line spectra were obtained for 20 of them, and for onethe VLA detection was not confirmed. Optical surface photometry has been carried out for 10 objects, for 8 of which opticalspectroscopy was obtained as well. Based on various criteria, we classify two sample galaxies (ce-143 and ne-204) as genuinedwarfs whereas the remaining sample objects are found to be intrinsically luminous galaxies. For those objects investigated inoptical wavelengths we determine properties similar to those of actively star-forming galaxies, and find that approximately onehalf of them has properties intermediate between those of dwarf galaxies and low-luminosity disc galaxies. No optical redshiftscould be determined for two of the galaxies (sw-103 and sw-194) and their physical association with the H clouds detected attheir positions therefore remains uncertain. A particularly interesting object in our sample is the Tidal Dwarf Galaxy candidatece-061 in the galaxy merger IC 1182.

Key words. galaxies: abundances – galaxies: dwarf – galaxies: clusters: general –galaxies: clusters: individual: Hercules cluster

1. Introduction

It is well known from numerous studies based on observa-tions and simulations that the environment plays a fundamen-tal role in the evolution of bright galaxies, via galaxy-galaxyinteractions and/or interactions of galaxies with the intergalac-tic medium. The present work is part of an ongoing study ofthe properties of H -selected galaxies in clusters, for whichthe results on dwarfs in the Hydra cluster have already been

Send offprint requests to: J. Iglesias-Paramo,e-mail: [email protected]� Appendix A and Figs. 5 to 12 are only available in electronic form

at http://www.edpsciences.org

published (Duc et al. 1999, 2001a; hereafter Papers I and II,respectively). The reasons for studying H -selected objectsare the sensitivity of H to the environment and the selec-tion of star-forming objects hosting H regions, for which di-rect measurements of the metallicity are fairly easy. The mostremarkable result is the existence of dwarfs with an oxygenabundance significantly higher than expected from theluminosity-metallicity relation for field dwarf galaxies.

In this paper we present single-dish 21 cm H -line as wellas optical imaging and spectroscopy observations of a sam-ple of galaxies in the Hercules cluster, selected from the ob-jects detected in the blind VLA H line survey of the clusterby Dickey (1997). Our main aims are to study the structural

Article published by EDP Sciences and available at http://www.aanda.org or http://dx.doi.org/10.1051/0004-6361:20030626

http://www.edpsciences.org/

http://www.aanda.org

http://dx.doi.org/10.1051/0004-6361:20030626

454 J. Iglesias-Paramo et al.: A study of H -selected galaxies in the Hercules cluster

and spectrophotometric properties of these gas-rich galaxies,search for dwarfs among them and identify possible differencesto the general galaxy population evolving within the cluster.

The Hercules supercluster is one of the most massive struc-tures in the nearby Universe (Freudling 1995). It appears tobe a collection of three clusters, gravitationally bound, but farfrom dynamical relaxation: Abell 2151, classified as richnessclass 2, and Abell 2147 and Abell 2152, both classified as rich-ness class 1 (Barmby & Huchra 1998). In a previous single-dish study of the H properties of a sample of galaxies in theHercules supercluster, Giovanelli et al. (1981) found a strongdeficiency in the H mass-to-optical luminosity ratio of galax-ies in the Abell 2147 cluster, while an almost normal, or mildlydeficient, ratio was found for the galaxies in Abell 2151, therichest and densest of the three clusters. We have assumed adistance of 150 Mpc to the Hercules supercluster, based onan average redshift of 11 050 km s−1 for the cluster spirals(see D97) and a Hubble constant of 75 km s−1 Mpc−1.

The paper is organized as follows: the sample selectionand H and optical observations are described in Sect. 2. Theoptical and H properties of the sample galaxies are describedin Sect. 3, and a discussion of the results is presented in Sect. 4.Comments on individual objects are given in the Appendix.

2. Observations

2.1. Sample selection

Our study of the Hercules cluster is aimed at H -selected galax-ies, as well as at reported H clouds without optical counter-parts on Digital Sky Survey (DSS) images, selected from theblind VLA 21 cm H line survey of the cluster by Dickey(1997, hereafter D97) and reobserved by us at Arecibo in orderto confirm the detections and to obtain H line profiles with abetter velocity resolution. Our failure to confirm the H cloudswithout optical counterparts indicates that these were spurious,as described in van Driel et al. (2003).

In the first part of our project we have concentrated onH -selected dwarf galaxies in the Hydra cluster (Papers Iand II) from the VLA H study by McMahon (1993), whichhas a velocity resolution of 42 km s−1, similar to that of theD97 survey. We based on H line widths at 20% of peak fluxdensity value, W20, smaller than 130 km s−1, as objects withsuch narrow lines are good dwarf candidates. Of the 20 selectedobjects only 4 were found not to be dwarfs.

We could not apply such an effective H line width cri-terion aimed at selecting dwarfs to the D97 data, however.As the Hercules cluster is about three times more distant(150 Mpc) than the Hydra cluster (45 Mpc), the H profilesare correspondingly weaker and their widths more uncertain.The D97 H mass detection limit of about 5 × 108 M� allowsthe detection of only the most gas-rich dwarf systems, whilewe estimate that the uncertainty in the FWHM line widths,W50, is about 100 km s−1 for the fainter H detections, follow-ing Fouque et al. (1990). We raised the cut-off value for W 50

to 270 km s−1 for the Hercules cluster, thereby excluding onlythe most massive, inclined spiral systems.

All four fields of the D97 VLA study (three of which –ne, ce and sw – are located in Abell 2151, while the forth –47 – is centred on Abell 2147) were covered in our Arecibosingle-dish H study. Of the total of 25 galaxies with opti-cal counterparts in the Digital Sky Survey (DSS) and show-ing line widths smaller than 270 km s−1 detected in the 4 fieldsby D97, we observed the following 18 in the H line at Arecibo:ne-112, ne-142, ne-178, ne-204, ne-208, ne-240, ce-042, ce-048, ce-060, ce-061, ce-143, ce-176, ce-200, sw-103, sw-222,47-138, 47-166 and 47-211. The 7 others were ruled out be-cause they are very likely face-on spirals: ne-169, ne-222, ne-264, ce-122, ce-166, sw-159 and 47-030. In addition, we in-cluded in our list four galaxies showing line widths larger than270 km s−1 but with only tentative optical counterparts on theDSS (ne-250, ne-398, sw-194 and 47-154) in order to confirmthe H detections and to verify whether the optical associationsare real or not. A summary of the observations obtained for oursurvey, together with the centre positions and W50 line widthsfrom D97, is given in Table 1.

For our optical imaging and spectroscopic observations,only the central (ce) and southern (sw) fields could be covered.

2.2. H I line observations

We made our H line observations of the 22 H -selected galax-ies in the Hercules cluster with the refurbished 305 m AreciboGregorian radio telescope in May and June 2002. For furthertechnical details on the observations and the data reduction werefer to van Driel et al. (2003).

The total net integration time (on+off) was on aver-age 70 min per source, depending on the line strength,from 40 min for the strongest lines to 110 min for the weak-est signals, in ne-398 and 47-154. The velocity coverage isabout 2500 km s−1, the velocity resolution is 1.3 km s−1, andthe telescope’s half power beam width (HPBW) at 21 cm is3.′4 × 3.′6. For the telescope’s pointing positions the centre co-ordinates of the VLA H sources as given in D97 were used(see Table 1 of D97). The data were reduced using IDL rou-tines developed at Arecibo Observatory. A first-order baselinewas then fitted to the data and the velocities were corrected tothe heliocentric system, using the optical convention. All datawere boxcar smoothed to a velocity resolution of 9.1 km s −1

for further analysis, while the data of ne-398 and 47-154 weresmoothed to 19.5 km s−1.

2.3. Optical observations

Our optical observations are limited to objects in the ce and swfields of the D97 survey. Of all 10 sample galaxies in these twofields we obtained deep CCD images, and low-to-medium res-olution spectra for the 8 among these with the brightest opticalcounterparts (see Table 1).

2.3.1. Optical imaging

B, V and i-band images were taken for most objects in ouroptical sample, except ce-143 and sw-222, for which only V

J. Iglesias-Paramo et al.: A study of H -selected galaxies in the Hercules cluster 455

Table 1. Observations.

Name RA Dec W50 H det. Photom. Optical

VLA VLA VLA Arecibo emission

(J2000.0) (km s−1) lines

ne-112 16 06 37.6 18 23 49 262 yes

ne-142 16 06 22.5 18 00 03 171 yes

ne-178 16 06 13.8 17 57 11 217 yes

ne-204 16 06 05.9 18 09 20 125 yes

ne-208 16 06 05.7 18 16 43 125 yes

ne-240 16 05 58.1 18 24 41 262 conf.

ne-250 16 05 52.2 18 27 58 307 yes

ne-398 16 04 18.0 18 14 06 352 weak

ce-042 16 06 00.1 17 45 54 216: yes B,V, i yes

ce-048 16 05 55.7 17 42 39 171 conf. B,V, i yes

ce-060 16 05 44.6 17 42 19 171 conf. B,V, i no

ce-061 16 05 41.1 17 48 00 216 conf B,V, i yes

ce-143 16 05 20.6 17 52 02 171 yes V, i yes

ce-176 16 05 09.9 17 51 20 171 yes B,V, i yes

ce-200 16 05 06.7 17 47 00 216 conf. B,V, i yes

sw-103 16 04 00.6 17 15 13 261 yes B,V, i

sw-194 16 01 07.1 17 20 16 395 – B,V, i

sw-222 16 03 05.8 17 10 14 261 yes V, i yes

47-138 16 00 17.5 15 53 15 216 –

47-154 16 02 26.6 15 57 36 372 no

47-166 16 02 16.2 16 04 41 261 yes

47-211 16 01 55.8 15 42 29 261 yes

Arecibo H data: conf. indicates a spectrum confused by nearby galaxies,– Objects for which no sensitive spectra could be obtained (see Sect. 3.1).

and i imaging could be obtained. Observations were carriedout with the Wide Field Camera attached to the prime focus ofthe 2.5 m Isaac Newton Telescope (INT) of the Observatoriodel Roque de los Muchachos, in Spain, on June 5, 1999 andApril 26, 2000. Both fields were observed under photometricconditions. The WFC consists of a science array of four thinnedAR coated EEV 4k× 2k devices, plus a fifth used for autoguid-ing. The pixel scale is 0.33 arcsec pixel−1, which gives a totalfield of view of about 34′ × 34′. Given the particular arrange-ment of the detectors, an area of about 11 ′ × 11′ is not usableat the top right corner of the field.

The accuracy of the photometry is about 0.10 mag. For thei-band frames, the accuracy is slightly poorer due to residualfringing and the uncertainties of the colour term δ(V−i) is about0.15 mag. Although a Sloan-Gunn i filter was used instead ofCousin I, the reported I magnitudes correspond to the Cousinsystem, as photometric standards from Landolt (1992) were ob-served and a linear relationship with a slope unity was foundbetween the expected number of counts for each of the fil-ters. The astrometry of the images was carried out using USNOguide stars. Detailed V-band images as well as (V − I) colourmaps are shown in Sect. 3.2.

Table 2. Journal of the optical spectroscopic observations.

Object Date Telescope PAa Widthb

ce-042 April 28 2001 NOT 5.6 1.0

ce-048 June 19 2000 WHT 336.4 1.0

ce-060 March 19 2001 NOT 319.3 1.0

ce-061 June 19 2000 WHT 96.6 1.1

ce-143 July 11 2000 WHT 51.5 1.0

ce-176 April 26 2000 WHT 317.5 1.1

ce-200 March 19 2001 NOT 70.7 1.0

sw-222 May 14 2001 NOT 81.0 1.0

a In degrees, from North towards East.b Width of the slit, in arcsecs.

2.3.2. Optical spectroscopy

Medium and low-resolution spectroscopy was carried out atthe 4.2 m William Herschel Telescope (WHT) and the 2.5 mNordic Optical Telescope (NOT) at the Observatorio del Roquede los Muchachos, in Spain. Table 2 shows the diary of the


spectroscopic observations. Observations at the WHT wereperformed during several nights using the double arm spectro-graph ISIS, with the dichroic splitting the beam set at 5700 Å.For most of the galaxies observed at the WHT, the CCD set-up was two 1k × 1k Tektronix per arm. The gratings used wereR300B and R316R, giving nominal dispersions of 1.54 Å pix −1

and 1.49 Å pix−1 for the blue and red arms, respectively.The corresponding spectral coverages were 3735–5311 Å and6118–7643 Å, respectively, and the spatial scale in both de-tectors was 0.36 arcsec pix−1. Note that ce-143 was observedwith a different set-up, using a EEV 2k × 4k detector on theblue arm, giving a nominal dispersion of 0.86 Å pix −1 and aspatial scale of 0.2 arcsec pix−1. In all cases, the slit widthwas set to match the seeing, about 1.′0 for most of the ob-jects. Observations at the NOT were taken using the faint ob-ject spectrograph ALFOSC, with a 2k × 2k LORAL detectorand Grism #4, giving a nominal dispersion of 3.3 Å pix −1. Thetotal wavelength coverage with this set-up was 3200–9100 Å,and the spatial scale was 0.18 arcsec pix−1.

The slit was always centred on the galaxy nucleus and po-sitioned along the position angle of the major axis, as listedin Table 2. Although spectro-photometric standard stars wereobserved for flux calibration, only the relative fluxes of theemission lines are reliable, since several nights were non-photometric.

The emission lines were measured with the SPLOT pack-age running on IRAF. For each emission line five independentmeasures were performed, and the adopted fluxes and errorsare, respectively, the average and standard deviation of the fivemeasures. For ce-042, ce-200 and sw-222 the intensities of the[N] doublet and the Hα line were determined by fitting thenon-resolved triplet with three Gaussians. Similarly, the fluxesof the [S] lines of sw-222 were obtained by fitting the dou-blet with two Gaussians. Larger errors resulted from this de-blending process. No correction was made for line absorptionin the Balmer lines. Only ce-200 shows an absorption featureat Hβ, for which we simply measured the intensity from thebase of the emission line. For some of the spectra obtained atthe WHT, we had to rescale the red part in order to get the samecontinuum levels at both sides of the gap, since it is well estab-lished that the continuum of H regions/galaxies is smooth atthese wavelengths. All lines were dereddened using the extinc-tion coefficient derived from the Balmer decrement Hα/Hβ andassuming a theoretical value for the intrinsic line ratio of 2.89(Brocklehurst 1971). No extinction correction was applied togalaxies for which the observed Hα/Hβ flux ratio was consis-tent with this value, within the errors.

3. Results

3.1. H I properties

The Arecibo H spectra of the 20 galaxies for which we couldobtain sensitive spectra are shown in Fig. 1, smoothed to aresolution of 9.1 km s−1 for most objects, and to 19.5 km s−1

for the weak line signal of ne-398 and for the undetected ob-ject 47-154. Not shown are the spectra of sw-194 and 47-138,

for which no sensitive H observations could be obtained dueto the proximity of strong continuum sources.

Besides the VLA data of D97, on which the present studyis based, published H detections (see Appendix A) were onlyfound for the merger system IC 1182, which contains the ce-061 tidal dwarf galaxy in one of its tails.

We compared (Table 3) the Arecibo global H profile pa-rameters to those of the D97 VLA observations, which havea synthesized beam size (HPBW) varying from 20 ′′ × 21′′to 26′′ × 29′′, a velocity resolution of 44.2 km s−1 (degradedto 88 km s−1 for the determination of the profile parameters),an rms noise of about 0.13 mJy/beam per channel map at thecentre of the primary beam, and a pixel size of 6 ′′ × 6′′, i.e.about 24 pixels per synthesized beam. Values in brackets in-dicate the 4 Arecibo profiles estimated to be significantly con-fused by line emission from nearby galaxies: ne-240, ce-048,ce-060 and ce-200 (see Appendix A for comments on these ob-jects). Listed in the columns are the following data; note thatall radial velocities in this paper, both optical and H , are in theheliocentric system, using the conventional optical definition(V = c(λ − λ0)/λ0): (1) the galaxy’s name, from D97, (2) thecentre velocity of the VLA profile, VHI, (3) the width of theVLA profile at 50% of the maximum flux density, W50, (4) theIH integrated VLA H line flux (see the description below),(5) the Iext integrated VLA H line flux (see the descriptionbelow), (6) the centre velocity of the Arecibo H profile, V HI,(7) the width of the Arecibo profile at 50% of the maximumflux density, W50, (8) the width of the Arecibo profile at 20%of the maximum flux density, W20, (9) the integrated AreciboH line flux, IHI and (10) the rms noise levels of the Arecibospectra.

As it is in principle not straightforward to determine the in-tegrated H line profiles parameters of faint objects from inter-ferometric data, four different methods were used in D97. Weconverted the H masses listed in D97 to integrated H linefluxes assuming the cluster distance of 110.5 Mpc adoptedin D97. The characteristics of these methods are as follows:IH is the line flux measured by integrating the spectra overthe group of contiguous pixels above threshold, I ext is the lineflux integrated over a larger area, estimated by statistical teststo contain the total line emission, while Ipeak and Iint are ob-tained by fitting a two-dimensional Gaussian to the velocity-integrated H column density map, where Ipeak corresponds tothe line flux within the central beam area and I int to the inte-grated flux of the Gaussian. The latter is notoriously unstable.

Although observations with a single dish telescope likeArecibo result in only one spectrum per pointing position andthe derivation of integrated H profile parameters is straightfor-ward, these profiles depend on the instrument’s beam pattern,which can lead to confusion with other objects in the beam (seeAppendix A), and single-dish data are more sensitive to RFIthan interferometric data.

We estimated the uncertainties, σVHI , in the central H ve-locities of the Arecibo line profiles, following Fouque et al.(1990):

σVHI = 4R0.5P0.5W X−1 [km s−1] (1)


Fig. 1. Arecibo 21 cm H line spectra of 20 galaxies in the Hercules cluster; for details on the ce-61 and ce-86 sources, see comments on ce-061(Appendix A). The velocity resolution of the data is 9.1 km s−1 for most spectra, and 19.4 km s−1 for ne-398 and 47-154.

where R is the velocity resolution in km s−1, PW =

(W20−W50)/2 in km s−1 and X is the signal-to-noise ratio of aspectrum, which we defined as the ratio of the peak flux den-sity and the rms noise. The estimated uncertainties vary from 2to 14 km s−1 and are on average 4.6± 2.5 km s−1. According toFouque et al., the uncertainty in the line widths is 2σVHI for W50

and 3σVHI for W20.

For the 13 objects for which an unambiguous comparisonbetween the integrated Arecibo and VLA line fluxes can bemade (see Table 1) the average ratio, and its σN deviation, be-tween the two VLA flux determinations made without fitting aGaussian to the source, IH and Iext, and the Arecibo flux, IHI, isIH = 0.66 ± 0.27 IHI and Iext = 1.03 ± 0.47IHI.


Table 3. Comparison of basic VLA and Arecibo H observational data.

———– VLA data ———— —————— Arecibo data ——————–

Name VHI W50 IH Iext VHI W50 W20 IHI rms

km s−1 km s−1 Jy km s−1 km s−1 km s−1 km s−1 Jy km s−1 mJy

ne-112 11 046 262 0.23 0.24 11 014 ± 6 156 188 0.49 ± 0.09 0.63

ne-142 11 711 171 0.19 0.36 11 726 ± 5 128 153 0.35 ± 0.07 0.47

ne-178 11 556 217 0.63 0.92 11 591 ± 4 191 242 1.10 ± 0.09 0.49

ne-204 11 467 125 0.56 0.70 11 451 ± 3 62 93 0.54 ± 0.06 0.60

ne-208 11 556 125 0.10 0.20 11 521 ± 7 218 244 0.25 ± 0.07 0.39

ne-240 11 645 262 0.18 0.32 (11 636 ± 6 233 241 0.21) 20.39

ne-250 11 645 307 0.88 1.32 11 686 ± 5 283 307 0.84 ± 0.08 0.39

ne-398 10 602 352 0.50 1.45 10 820 ± 14 364 377 0.15 ± 0.06 0.22

ce-042 11 919: 216: 0.46: 0.67: 11 910 ± 3 127 135 0.33 ± 0.07 0.54

ce-048 11 145 171 0.20 0.36 (11 134 ± 2 208 211 0.53) 0. 1

ce-060 11 100 171 0.27 0.41 (11 085 ± 4 130 150 0.51) 0. 9

ce-061 10 104 216 3.16 3.44 10 263 ± 3 595 666 3.44 ± 0.18 0.76

ce-143 11 587 171 0.16 0.31 11 603 ± 3 95 101 0.14 ± 0.04 0.40

ce-176 9905 171 0.43 0.51 9919 ± 4 150 183 0.65 ± 0.08 0.56

ce-200 9927 216 0.10 0.17 (9938 ± 4: 177: 189 0.46:) 0.43

sw-103 11 021 261 0.15 0.24 10 999 ± 4 155 159 0.18 ± 0.06 0.38

sw-222 10 093 261 0.25 0.49 10 076 ± 4 192 226 0.79 ± 0.09 0.50

47-154 10 633 372 0.12 0.14 – 0.32

47-166 9849 261 0.45 0.62 9887 ± 3 248 293 0.84 ± 0.06 0.35

47-211 10 423 261 0.30 0.58 10 460 ± 3 280 295 0.83 ± 0.10 0.51

Arecibo H line parameters in brackets are for objects with profiles confused by nearby objects.

We also estimated the uncertainties, σIHI , in the integratedline fluxes of the Arecibo line profiles following Fouque et al.(1990), assuming that this formula, which was developed forNancay data, can be applied to Arecibo data as well:

σIHI = 5R0.5I0.5HI h0.5X−1 [Jy km s−1] (2)

where R and X are as in Eq. (1) and h is the peak flux densityof the profile, in Jy. For the 13 abovementioned galaxies theestimated uncertainties vary between 5 and 30% of IHI and areon average (0.16 ± 0.08) × IHI.

A comparison of the W50 line widths measured with theVLA, which were used for the galaxy selection, and at Arecibo(Fig. 2) clearly shows the effect of the difference in velocityresolution, 88 and 9 km s−1, respectively. It appears that, onaverage, the VLA line widths have been overestimated by abouthalf a VLA channel width, i.e. 20 km s−1.

Table 4 shows a summary of the H properties of the sam-ple galaxies as well as the absolute B magnitudes of their op-tical counterparts. Listed in the columns of this Table are thefollowing data: (1) the galaxy name, from D97, (2) the centralradial velocity of the H profile, VHI, (3) the global form ofthe Arecibo spectra that are not confused, where DH denotesa double-horned profile, FT a flat topped one, G a Gaussianone and LS a lopsided profile, (4) the total H mass, MHI,(5) the relative H gas content, MHI/LB, (6) the FWHM of theH line, W50, and (7) the absolute magnitude in the B band,MB, from NED. For the H -related properties we preferentially

Fig. 2. Comparison of W50 line widths of integrated H profiles mea-sured at Arecibo and the VLA, see the text for details. To guide theeye, a line with slope of unity has been plotted; it does not represent afit to the data.

used our Arecibo spectra, except for cases of confusion withnearby galaxies or a nearby strong continuum source, wherewe adopted the D97 VLA data, using the Iext estimate for theintegrated line flux.


Table 4. Basic properties of the sample galaxies.

Ident. VHI Spec. MHI MHI/LB W50 MB

km s−1 forma 109 M� M�/L�,B km s−1 mag

ne-112 11 014 LS 1.7 0.24 156 −19.1

ne-142 11 726 FT 1.8 3.7 128 −16.2

ne-178 11 591 DH 5.8 0.56 191 −19.5

ne-204 11 451 G 2.9 6.8 62 −16.0

ne-208 11 521 G 1.3 1.1 218 −17.1

ne-240 (11 645) – (1.7) (1.5) (262) −18.2

ne-250 11 686 FT 4.4 0.96 283 −19.8

ne-398 – – – – – −16.3

ce-042 11 910 DH 1.7 0.51 127 −17.2

ce-048 (11 145) – (1.9) (3.5) (171) −15.4

ce-060 (11 100) – (2.2) (2.1) (171) –

ce-061 (10 104) – (18.2) (7.3) (216) –

ce-143 11 603 G 0.74 0.35: 95 −18.0

ce-176 9919 LS 3.4 1.0 150 −18.3

ce-200 (9938) – (0.90) (0.43) (216) −17.6

sw-103 10 999 DH 0.95 1.1 155 −16.8

sw-194 (10 159) – (2.0) 15.7: (395) −14.8

sw-222 10076 G 4.2 0.35: 192 −19.8

47-138 (11 284) – (1.2) 0.30 (216) −18.4

47-154 (10 633) – (0.74) – (372) –

47-166 9887 DH 4.4 1.4 248 −15.0

47-211 10 460 DH 4.4 0.41 280 −19.8

a The H spectrum forms are: DH double horned, FT flat topped, G Gaussian, and LS lop-sided.

As mentioned above, the reported VLA H detections oftwo of the galaxies – ne-398 and 47-154 – were not confirmedat Arecibo.

3.2. Optical properties

In this section we study the results obtained for the sub-sample for which we obtained optical spectra and/or imag-ing. Figures 3a, b show our V-band images of the ce andsw fields, respectively, with superimposed contours showingthe H clouds detected in D97. Figure 4 displays the opticalspectra of all galaxies with emission lines and Table 6 liststheir spectro-photometric data. Uncertainties derived from thedeblending process for the [N] and [S] lines of the galaxiesobserved at the NOT are not included in the table.

3.2.1. Structural properties

In Table 5 we list the absolute magnitudes and colours obtainedfor the 10 objects of the optical subsample, including the twotentative detections sw-103 and sw-194. For the seven galaxieswhich show optical emission lines, the magnitudes are extinc-tion corrected using the C(Hβ) values listed in Table 6. The

magnitudes listed in this table will be used hereafter for thissubsample of galaxies.

For the surface photometry analysis we employed im-proved versions of the techniques described in Papaderoset al. (1996). Before determining the profiles we removedforeground stars and background galaxies intersecting theH galaxies and smoothed all images of a given object to thesame resolution. Surface brightness profiles (SBPs) were cor-rected for extinction inside the galaxies using the C(Hβ) valueslisted in Table 6.

In Table 7 we summarize the derived photometric proper-ties. Listed in the 12 columns of this table are the followingdata: (1) the parameters b and q, describing, respectively, theintensity distribution of the latter fitting formula near the cen-tre, (3 and 4) the central surface brightness, µE,0, and exponen-tial scale-length α, respectively, of the Low Surface Brightness(LSB) component, as obtained from linear fits to the exponen-tial regime of each profile and weighted by the photometricuncertainty of each point – note that Eq. (22) in Papaderoset al. (1996) predicts for the LSB component an actual cen-tral surface brightness 2.5 log (1 − q) mag fainter than the ex-trapolated value µE,0, listed in Col. 3, (5) the corresponding to-tal magnitude of the LSB component as obtained for a pure


Fig. 3. a) VLA H column density contours superposed on our optical V-band image of field ce. The labels indicate the H -selected galaxies inthis field. RA and Dec are in J2000. The H cloud ce-070 reported by van Driel et al. (2003) is also labeled.

Fig. 3. b) VLA H column density contours superposed on our optical V-band image of field sw. The labels indicate the H -selected galaxiesin this field. RA and Dec are in J2000. The H clouds sw-089, sw-146 and sw-201 reported by van Driel et al. (2003) are also labeled.

or a modified exponential distribution, (6–9) list quantities ob-tained from profile decompositions; in (8) and (6) are listed,respectively, the isophotal radii E25 and P25 of the LSB com-ponent and of the luminosity component in excess of it, bothdetermined at the 25 mag arcsec−2 level, while in (9) and (7)are listed, respectively, the corresponding apparent magnitudeof these components, mE25 and mP25, (10) the total apparentmagnitude, as inferred from the SBP integration out to the lastpoint, mSBP, (11) the effective radius, reff , and the radius r80,which encircles 80% of the galaxy’s total flux, and (12) theSersic index η resulting from fitting Eq. (5) in Papaderos et al.(1996) to the BSP.

V-band images, (V − I) colour maps, surface brightnessand colour profiles of the galaxies are presented in Figs. 5

to 12. A common property of all selected galaxies is the ab-sence of notable colour gradients. In all cases these do notexceed 0.1 mag kpc−1, as also reported for many dIs (e.g.Patterson & Thuan 1996; van Zee 2001). The situation is strik-ingly different in the inner regions (R∗ <∼ E25) of BCDs, wherecolour gradients of up to γ+ ∼ 1.8 mag kpc−1 have been ob-served, like H1034-2558 in the Hydra cluster (Paper I) andother examples in Papaderos et al. (1996a), Doublier et al.(1999) and Cairos et al. (2001).

For all galaxies except ce-061, an exponential fitting lawprovides a reasonable approximation to the SBPs in their outerlow surface brightness regime. For half of the sample galax-ies, however, inwards extrapolation of this outer exponen-tial slope predicts a higher intensity than actually observed.


Fig. 4. Optical spectra of the seven galaxies showing emission lines.


Table 5. Absolute magnitudes and colours.

Name MV (B − V) (V − I)

ce-042 −18.53 0.17 0.29

ce-048 −16.60 0.28 0.55

ce-060 −17.24 0.23 1.06

ce-061 −18.24 0.23 0.00

ce-143 −18.13† – 0.51

ce-176 −18.72 0.44 0.65

ce-200 −18.12 0.34 0.74

sw-103 −16.74 – –

sw-194 −15.24 – 1.80

sw-222 −20.02† – 0.68

† The average (B − V) = 0.3 was assumed to estimate MB.

This type of convex profile with an exponential distributionin the outer parts and levelling off in the inner part (within1–3 disc scale lengths) is not rare among low-luminosity dwarfellipticals (e.g. Binggeli & Cameron 1991), dwarf irregulars(Patterson & Thuan 1996; van Zee 2000), blue LSB galax-ies (Ronnback & Bergvall 1994), near-infrared selected LSBs(Monnier Ragaigne et al. 2003), and has been reported in afew blue compact dwarfs (e.g., Fricke et al. 2001; Guseva et al.2001; Vennik et al. 2000). Note that similar SBPs have alsobeen derived for 4 of the H -selected dwarfs in the Hydra clus-ter (H1031-2818, H1031-2632, H1032-2722 and H1033-2642;see Paper I). For those systems, following the procedure out-lined in Guseva et al. (2001), we modelled the LSB componentusing Eq. (22) in Papaderos et al. (1996).

In Fig. 13 we show MB as function of µ0,B for the Herculesand Hydra cluster galaxies. We used the average (B−V) colourof 0.3 of the other galaxies in the subsample for ce-143and sw-222, and included them in the plot. Also shown are theloci occupied by spiral discs, ellipticals and spiral bulges, dwarfIrregulars and dwarf ellipticals from Binggeli (1994), as wellas the loci occupied by the LSB galaxies from van der Hulst(1998). All Hydra cluster galaxies and 5 of the 8 Hercules ob-jects lie in the zone occupied by the faintest disc-like galaxiesand by dwarf systems.

3.2.2. Star formation activity

Not surprisingly for an H -selected sample, seven objects outof the eight for which we took optical spectra show emissionlines, with line ratios typical of H regions, indicating theyare star-forming objects. The only quiescent galaxy is ce-060,which shows no sign of ongoing star formation. Unfortunately,non-photometric weather conditions did not allow us to derivea reliable star-formation rate from the Hα luminosity.

In Fig. 14 we show the absolute blue magnitudes M B asfunction of the Hβ equivalent widths for the Hercules andHydra cluster galaxies. For comparison we have added theemission line galaxies from Salzer et al. (1989) as well asthe upper envelope from Vılchez (1995). As can be seen from

the figure, the Hercules cluster galaxies have Hβ equivalentwidths normal for their B-band luminosities, compared to theSalzer et al. sample, whereas the Hydra cluster dwarfs seemto delineate the lower envelope of the locus occupied by theSalzer et al. sample.

Broad-band colours can also be used as a diagnostic ofstar formation, and we show the (B − V) vs. (V − I) diagramfor our sample galaxies in Fig. 15. For comparison, we haveadded the ellipticals from Goudfrooij et al. (1994), the spi-rals from Heraudeau & Simien (1996), the amorphous galax-ies from Gallagher & Hunter (1987) and the nearby dwarfsfrom Makarova (1999) – the latter objects were defined bySandage & Brucato (1979) as blue galaxies showing centrallyconcentrated light distributions, with a high H gas content,low rotation velocities and heavy-element abundances typicalof Magellanic Irregulars. We notice that all 7 of our Herculescluster galaxies with both colours available are concentrated to-wards the bottom left region of the diagram, i.e. the blue part ofthe sequence delineated by the amorphous and dwarf samples,where no ellipticals and only a marginal number of spirals arefound.

3.2.3. Metallicity

The H -based selection process resulted in star–forming galax-ies hosting H regions of which the metallicity of the ionisedgas could be estimated from their emission lines. As the[O]λ4363 Å line was not detected for any of the galaxies,the oxygen abundances were computed from the [O]λ3727 Åand [O]λ4959,5007 Å emission line fluxes, using severalabundance calibrations: the R23 method calibrated theoreticallyby McGaugh et al. (1991), in its parameterised form fromKobulnicky et al. (1999), the more recent p-method (Pilyugin2000, 2001) and the independent method of van Zee et al.(1998), based on the [N]/Hα line ratio. The first two meth-ods each provide two possible values of the metallicity for agiven observable. To resolve the degeneracy and choose be-tween the lower and upper values, we relied on the strength ofthe [N]/[O] ratio (see van Zee et al. 1998). The values of12 + log (O/H) obtained with these three methods, as well asthe final adopted values, are listed in Table 8.

The criteria used to choose the final adopted abundanceslisted in Col. 7 were the following: the abundances were prefer-ably derived using the p-method, for consistency with previouswork in this program (Paper II). Although the method basedon [N]/Hα is not very reliable, because of blending problemsof the Hα and [N] lines in the low resolution NOT spectra,it was preferred for those objects (ce-042 and ce-200, markedwith a colon in Table 8) whose observed line fluxes are notwithin the validity range of the p-method, and which yield in-consistent metallicities: i.e. the oxygen abundance read on thelower branch is larger than that read on the upper branch.

The median oxygen abundance of these 7 Hercules clus-ter sample galaxies is about one third solar (12 + log (O/H) =8.27 ± 0.31), with a quite large scatter. Figure 16 shows theirmetallicities as function of absolute B magnitude; the typicaluncertainty in the oxygen abundances is indicated by the error


Table 6. Extinction-corrected emission line flux ratios with respect to Hβ†.

ce-042 ce-048 ce-061 ce-176 ce-200 ce-143 sw-222

[O] λ3727 Å 6.149 0.932 2.475 3.062 6.234 3.505 3.044

(1.108) (0.050) (0.122) (0.343) (0.956) (0.322) (0.154)

Hβ λ4861 Å 1.000 1.000 1.000 1.000 1.000 1.000 1.000

(0.071) (0.012) (0.016) (0.052) (0.042) (0.037) (0.016)

[O] λ4959 Å – 1.424 0.940 1.004 – 0.834 0.597

– (0.040) (0.033) (0.086) – (0.065) (0.038)

[O] λ5007 Å 2.955 3.559 2.753 2.287 1.878 2.730 1.867

(0.299) (0.078) (0.069) (0.154) (0.135) (0.142) (0.054)

[N] λ6548 Å 0.203 – 0.142 – 0.086 – 0.167

(0.080) – (0.009) – (0.063) – (0.015)

Hα λ6563 Å 2.891 2.756 2.891 2.770 2.890 2.889 2.890

(0.528) (0.108) (0.105) (0.334) (0.344) (0.277) (0.109)

[N] λ6584 Å 0.466 – 0.332 0.298 0.186 0.447 0.407

(0.116) – (0.021) (0.045) (0.068) (0.067) (0.019)

[S] λ6717 Å – – 0.402 0.529 – 0.703 0.573

– – (0.022) (0.066) – (0.096) (0.026)

[S] λ6731 Å – – 0.300 0.365 – 0.513 0.371

– – (0.017) (0.047) – (0.057) (0.017)

Vhel (km s−1) 11 840 11 181 10 012 9875 9940 11 616 10 087

(90) (39) (76) (95) (65) (63) (20)

C(Hβ) 0.298 0.000 0.004 0.000 0.103 0.116 0.166

(0.117) (0.025) (0.023) (0.077) (0.076) (0.061) (0.024)

EW(Hβ) (Å) 10.4 47.7 30.9 14.3 8.9 13.0 14.5

(1.0) (2.5) (0.9) (1.1) (0.4) (0.6) (0.3)

† Quantities in parenthesis correspond to the errors of the quantities quoted above.

bar. For the two galaxies without an available B magnitude,ce-143 and sw-222, a (B − V) colour of 0.30 – the medianvalue measured for the other Hercules cluster galaxies – was as-sumed. For comparison, we added the isolated dwarf Irregularsof Richer & McCall (1995), the LSBs from van der Hulstet al. (1998), the H -selected dwarfs from the Hydra cluster(Paper II) and the tidal dwarf galaxies (TDGs) from Duc et al.(2000). The straight line represents the empirical relationshipfound by Richer & McCall (1995) for their dwarf sample.

The majority of the Hercules cluster galaxies show metal-licities following the empirical relationship for Irregulardwarfs, taking into account the uncertainties. Exceptions arece-042, which looks overabundant for its luminosity and lieswell within the loci of the LSBs and the TDGs, and ce-048,with a metallicity lower than expected for its luminosity andwhose spectrum resembles that of typical BCDs. The presenceof a few over–metallic object was also reported in the Hydracluster by Duc et al. (2001a). Unique among our objects in theHercules cluster is the TDG ce-061 (Braine et al. 2001), whosedynamical nature as a gravitationally bound system inside thetidal tail of the IC 1182 merger system has been confirmed

recently through Hα line Fabry-Perot imaging (Duc & Amram2003).

3.2.4. The two tentative optical detections

Figures 17a, b show V-band contour plots of the two opticaldetections we tentatively associate with the VLA H sourcessw-103 and sw-194. The filled circles indicate the centre posi-tions of the H clouds. Although in both cases the differencein position between the centres of the optical and H sourcesare small compared to the VLA beam size (see below), whichfavours their association, the physical link between the H de-tections and the optical counterparts could not be proven un-equivocally, as we could not obtain optical spectra for thesetwo galaxies since they are too faint. Such optical spectra areneeded to confirm their associations to the H clouds.

Unfortunately, sw-103 is too close to a bright star fora detailed morphological analysis. The difference betweenthe optical and H centre is only about 4 ′′, or one fifth ofthe VLA HPBW. Although its V-band magnitude could be


Table 7. Photometric results and structural properties of the galaxies with definite optical counterpartsa.

Name Band µ0 α mfitLSB P25 mP25 E25 mE25 mSBP reff ,r80 η

mag/��′′ kpc mag kpc mag kpc mag mag kpc

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

ce-042 B 21.25 ± 0.50 2.14 ± 0.23 17.67 2.96 19.35 7.14 18.06 17.50 4.35, 7.29 0.91

(2.8,0.9) V 21.26 ± 0.45 2.24 ± 0.22 17.59 3.48 18.91 7.87 17.98 17.31 4.35, 7.36 0.91

I 21.13 ± 0.52 2.32 ± 0.30 17.38 4.01 18.43 8.20 17.62 17.04 4.26, 7.31 0.85

ce-048 B 22.68 ± 0.16 1.17 ± 0.06 19.61 1.06 22.25 2.49 20.11 19.52 1.65, 3.06 0.89

� V 22.59 ± 0.15 1.19 ± 0.06 19.48 1.55 20.79 2.64 19.94 19.21 1.48, 2.87 0.77

I 22.51 ± 0.34 1.23 ± 0.10 19.32 1.89 19.68 2.83 19.76 18.72 1.30, 2.56 0.62

ce-060 B 23.42 ± 0.13 2.45 ± 0.13 18.74 0.89 23.00 3.56 19.67 18.85 3.26, 5.38 0.90

� V 23.15 ± 0.10 2.41 ± 0.10 18.51 1.41 19.14 4.10 19.25 18.59 3.07, 5.21 0.87

I 22.11 ± 0.13 2.44 ± 0.14 17.45 6.56 17.73 6.48 17.77 17.54 3.08, 5.23 0.83

ce-061 B 22.68 ± 0.27 2.41 ± 0.19 18.72 3.87 18.46 <P25 – 17.84 2.70, 5.33 0.63

(2.4,0.88) V 22.44 ± 0.23 2.38 ± 0.16 18.50 3.95 18.19 3.91 19.96 17.60 2.60, 5.21 0.61

I 21.50 ± 0.42 2.19 ± 0.25 17.72 3.91 17.80 6.93 18.12 17.06 2.69, 5.52 0.63

ce-143 V 21.32 ± 0.11 1.20 ± 0.04 18.2 – – 4.08 18.36 18.32 2.07, 3.48 1.09

� I 20.97 ± 0.13 1.24 ± 0.05 17.78 – – 4.59 17.91 17.82 2.03, 3.42 0.99

ce-176 B 21.89 ± 0.20 2.26 ± 0.10 17.69 2.01 20.39 6.45 18.06 17.54 4.22, 6.87 1.17

(2.1,0.7) V 21.61 ± 0.17 2.35 ± 0.10 17.33 3.16 19.22 7.32 17.62 17.12 4.11, 6.83 1.09

I 21.11 ± 0.31 2.41 ± 0.18 16.77 4.90 17.99 8.63 16.96 16.46 3.96, 6.71 1.00

ce-200 B 20.35 ± 0.32 1.16 ± 0.07 18.31 2.26 19.89 4.94 18.52 18.07 2.59, 4.30 1.41

(3.3,0.92) V 20.06 ± 0.21 1.16 ± 0.05 18.00 2.44 19.44 5.30 18.12 17.72 2.59, 4.30 1.39

I 19.35 ± 0.09 1.18 ± 0.03 17.27 2.67 18.67 6.24 17.36 17.00 2.59, 4.37 1.23

sw-222 V 20.31 ± 0.07 1.64 ± 0.03 16.51 – – 7.08 16.59 16.42 2.55, 4.40 0.97

� I 19.65 ± 0.08 1.65 ± 0.04 15.83 – – 8.13 15.88 15.74 2.53, 4.42 0.91

a: All listed values are corrected for extinction assuming a C(Hβ) as listed in Table 6.

estimated (mV ≈ 19.1), this was not possible in the I-band dueto the close bright star.

The other tentative case, sw-194, may well be spurious. Itis noted in D97 that this VLA H source could be an artefactdue to a nearby strong continuum source (see Appendix A).It is clearly extended (about 1.′5 × 0.′5) and four galaxies (seeFig. 17b) lie within its contour. Our tentative identificationconcerns the brightest and largest of the four objects only,which lies about 11′′ from the centre of the VLA source. Itis a very faint galaxy, for which we estimated mV ≈ 20.6 and(V − I) ≈ 1.8, which is redder than expected for an H -richgalaxy. Although the three other objects are fainter and smaller,their association with the H source remains possible.

4. Discussion

As mentioned earlier, one of our aims is the identificationof H -selected dwarf galaxies in clusters. Although differentdwarf selection criteria based on the H and optical properties

of galaxies are used in the literature, there is no precise anduniversally accepted definition of what exactly a dwarf galaxyis (e.g., Binggeli 1994). The classification criteria we can ap-ply to our data are: narrow H line width, Gaussian H lineshape, faint optical luminosity and central surface brightnessas a function of luminosity. Two galaxies, ne-398 and 47-154,could not be classified, as they were not detected in H atArecibo and therefore lack an accurate line width.

Despite having an optical morphology and structural pa-rameters reminiscent of dwarf or low-luminosity galaxies, mostof the H -selected objects in our sample have rather largeH line widths and sometimes double-horn profile shapes thatare inconsistent with them being low-mass objects.

The only two systems in our Hercules cluster sample thatsatisfy all four dwarf selection criteria are ne-204 and ce-143:both have a W50 profile width smaller than 100 km s−1, aGaussian H profile shape, an absolute blue luminosity fainterthan−18 (see Table 4) and they lie among the dwarfs in Fig. 13.


��

��

��

��

LSBLSB

Es + b

ulge

s?

D W A R F sdisks

dEIrr

(?)

nuclei

globulars

ss

s s

s

s sss

ss

ss

ss

ss

s

sMalin 1

ss

Hercules

Hydra

M32+

++

+

++

+

+

++ +

++

+++

++

+

+

++

+

Fig. 13. Comparison of the absolute blue magnitude and blue central surface brightness of different types of galaxies and stellar subsystems,see text for details. To the original version of Binggeli (1994) the loci of the LSB galaxies from van der Hulst (1998) have been added, as wellas the objects with available CCD surface photometry from our samples of H -selected galaxies in the Hercules and Hydra clusters.

Fig. 14. Absolute blue magnitude as function of Hβ equivalent width for the Hercules and Hydra cluster samples. For comparison we haveadded the emission line galaxies from Salzer et al. (1989), represented by small crosses. The straight line corresponds to the upper envelopefrom Vılchez (1995).


Fig. 15. (B − V) vs. (V − I) diagram for the Hercules cluster galaxies. Also shown are ellipticals from Goudfrooij et al. (1994), spirals fromHeraudeau & Simien (1996), amorphous galaxies from Gallagher & Hunter (1987) and nearby dwarfs from Makarova (1999).

Fig. 16. Metallicity vs. luminosity relationship for the Hercules and Hydra cluster galaxies, represented by open diamonds and triangles respec-tively. The typical uncertainty in the metallicity determinations is represented by the error bar at the bottom left. Also shown are the sample ofdwarfs from Richer & McCall (1995), the LSBs from van der Hulst et al. (1998) and the tidal dwarfs from Duc et al. (2000), represented bycrosses, squares and asterisks, respectively.


sw−103

Dec

.(J2

000)

R.A.(J2000)

Fig. 17. a) V-band contour plot of the tentative detection corresponding to sw-103. Contour levels are 25.5, 24.5, 23.5, 22.5and 21.5 mag arcsecs−2. The filled circle shows the centre position of the VLA H source from Dickey (1997).

sw−194

Dec

.(J2

000)

R.A.(J2000)

Fig. 17. b) V-band contour plot of the tentative detection corresponding to sw-194; we have assumed that the gas cloud is associatedwith the brightest galaxy in the field, 11′′ W of the centre position (see Sects. 3.2.4 and 4). Contour levels are 25.5, 24.5, 23.5, 22.5and 21.5 mag arcsecs−2. The filled circle shows the position of the VLA H source from Dickey (1997).

The levels of star formation in the four galaxies for whichoptical imaging and spectroscopy was obtained – as indicatedby their Hβ equivalent widths and optical colours – were foundto be typical of those of active star-forming galaxies.

The metallicities of two objects do not follow the sametrend as the five others: ce-042 was found to be overabundantfor its luminosity and to lie well within the loci occupied by

tidal dwarfs and LSBs, whereas ce-048 shows a very low metal-licity for its luminosity. In fact, it is likely that the presence ofa strong burst of star formation in ce-48, as its large Hβ equiv-alent width suggests, has raised its luminosity, thus shifting itsignificantly in the B-band luminosity vs. metallicity diagram.

The presence of two gas clouds with tentative optical de-tections (sw-103 and sw-194) should also be noted. While the


Table 8. Oxygen abundances (12 + logO/H).

Name p-method R23 N2Ha Adopted

low up low up abundance

ce-042 8.52 7.91 8.49 8.24 8.68 8.68:

ce-048 7.58 8.61 7.73 8.71 – 7.58

ce-061 7.88 8.41 7.98 8.62 8.53 8.41

ce-176 8.02 8.35 8.03 8.60 8.48 8.35

ce-200 8.66 7.94 8.43 8.31 8.27 8.27:

ce-143 8.10 8.26 8.14 8.51 8.66 8.26

sw-222 8.03 8.37 7.99 8.64 8.62 8.37

former was confirmed in H at Arecibo, the latter, which couldnot be observed at Arecibo, could be a spurious VLA detection,according to D97. If the optical counterparts are really physi-cally associated to the neutral gas clouds, they would be dwarfgalaxies according to their optical magnitudes of M B ∼ −16.9and −14.8, respectively. If our tentative identification of a sin-gle galaxy with sw-194 were correct and the H source notspurious (see Sect. 3.2.4), then its H mass-to-blue luminosityratio would be about 16 M�/L�,B, an unprecedented high valuefor a dwarf, while sw-103 would have a quite mundane ratioof 1.1 M�/L�,B. If the physical association between these H and optical sources were not confirmed, we may be dealingwith gas clouds without detected optical counterparts in a clus-ter, whose existence would impose constraints on the clusterproperties.

Our accumulated data on the Hercules and Hydra clustergalaxies will be discussed further in a future paper on the influ-ence of the cluster environment on galaxies, specifically dwarfs(Duc et al., in preparation).

5. Conclusions

According to various diagnostics, two of our sample galaxies(ce-143 and ne-204) can be classified as dwarfs. Of the galax-ies studied in the optical, the star formation properties are sim-ilar to those of other samples of actively star forming galaxiesand their metallicities are consistent with the blue luminosity-metallicity relation of nearby dwarfs, except for two galaxies,one of which is over-metallic and the other under-metallic forits luminosity. For two others the physical association with theH clouds seen superimposed on them could not be proven,since no optical redshifts could be obtained. If confirmed, oneof them (sw-194) would have an extremely high H content.A particularly interesting object in our sample is located in theIC 1182 merger system: the Tidal Dwarf Galaxy ce-061, whichshows a rather high luminosity, metallicity and H content.

Acknowledgements. Some of the data presented here have been takenusing ALFOSC, which is owned by the Instituto de Astrofısica deAndalucıa (IAA) and operated at the Nordic Optical Telescope un-der agreement between IAA and the NBIfAFG of the AstronomicalObservatory of Copenhagen. The NOT and WHT are operated on theisland of La Palma by the NOT and ING groups, at the Observatoriodel Roque de Los Muchachos of the Instituto de Astrofısica de

Canarias. We also acknowledge the Service Program of the ING groupfor useful observations presented in this paper. We would like tothank the staff of Arecibo Observatory for their help with the obser-vations and data reduction, especially P. Perrillat. IRAF, the ImageReduction and Analysis Facility, is written and supported at theNational Optical Astronomy Observatory. We have made use of theNASA/IPAC Extragalactic Database (NED), which is operated by theJet Propulsion Laboratory, California Institute of Technology, undercontract with the National Aeronautics and Space Administration, andthe Lyon-Meudon Extragalactic Database (LEDA). JIP acknowledgesthe Fifth Framework Program of the European Union for a MarieCurie Postdoctoral Fellowship. Research by PP has been supportedby the Deutsches Zentrum fur Luft– und Raumfahrt e.V. (DLR) undergrant 50 OR 9907 7. C.B.F. WvD acknowledges the CNRS-SDU fortravel support through the ASTE program.

References

Barmby, P., & Huchra, J. P. 1998, AJ, 115, 6Bieging, J. H., & Biermann, P. 1983, AJ, 88, 161Binggeli, B., & Cameron, L. M. 1991, A&A, 252, 27Binggeli, B. 1994, A note on the Definition and Nomenclature

of Dwarf Galaxies, in Proc. ESO/OHP Workshop on DwarfGalaxies, ed. G. Meylan, & P. Prugniel (ESO, Garching), 13

Bird, C. M., Dickey, J. M., & Salpeter, E. E. 1993, ApJ, 404, 81Bothun, G. D., Stauffer, J. R., & Schommer, R. A. 1981, ApJ, 247, 42Bothun, G. D., Aaronson, M., Schommer, R. A., et al. 1985, ApJS, 57,

423Braine, J., Duc, P.-A., Lisenfeld, U., et al. 2001, A&A, 378, 51Brocklehurst, M. 1971, MNRAS, 153, 471Cairos, L. M., Vılchez, J. M., Gonzalez-Perez, J. N., Iglesias-Paramo,

J., & Caon, N. 2001, ApJS, 133, 321Cayatte, V., van Gorkom, J. H., Balkowski, C., & Kotanyi, C. 1990,

AJ, 100, 604Dickey, J. M. 1997, AJ, 113, 1939Doublier, V., Caulet, A., & Comte, G. 1999, A&AS, 138, 213Duc, P.-A., & Mirabel, I. F. 1994, A&A, 289, 83Duc, P.-A., & Mirabel, I. F. 1998, A&A, 333, 813Duc, P.-A., Papaderos, P., Balkowski, C., et al. 1999, A&AS, 136, 539

(Paper I)Duc, P.-A., Brinks, E., Springel, V., et al. 2000, AJ, 120, 1238Duc, P.-A., Cayatte, V., Balkowski, C., et al. 2001a, A&A, 369, 763

(Paper II)Duc, P.-A., Balkowski, C., Cayatte, V., et al. 2001b, Environmental

Effects on H -rich Dwarf Galaxies in the Hydra and HerculesCluster, in Dwarf Galaxies and their Environment, ed. K. S.de Boer, R.-J. Dettmar, & U. Klein, 153

Duc, P.-A., & Amram, P. 2003, in preparationFerguson, H. C., & Binggeli, B. 1994, A&ARv, 6, 67Fouque, P., Bottinelli, L., Durand, N., Gouguenheim, L., & Paturel, G.

1990, A&AS, 86, 473Freudling, W. 1995, A&AS, 112, 429Fricke, K. J., Izotov, Y. I., Papaderos, P., Guseva, N. G., & Thuan,

T. X. 2001, AJ, 121, 169Gallagher, J. S., & Hunter, D. A. 1987, AJ, 94, 43Giovanelli, R., Haynes, M. P., & Chincarini, G. L. 1981, ApJ, 247,

383Goudfrooij, P., Hansen, L., Jorgensen, H. E., et al. 1994, A&AS, 104,

179Guseva, N. G., Izotov, Y. I., Papaderos, P., et al. 2001, A&A, 378, 756Hoffman, G. L., Helou, G., & Salpeter, E. E. 1988, ApJ, 324, 75


Heraudeau, P., & Simien, F. 1996, A&AS, 118, 111Hopp, U., Kuhn, B., Thiele, U., et al. 1995, A&AS, 109, 537Huang, Z., & Sarazin, C. L. 1996, ApJ, 461, 622Huchra, J., Geller, M. J., Clemens, C. M., Tokarz, S. P., & Michel,

A. 1995, The CfA Redshift Catalog (Cambridge: Harvard-Smithsonian Center for Astrophysics)

Kobulnicky, H. A., Kennicutt, R. C., & Pizagno, J. L. 1999, ApJ, 514,544

Landolt, A. U. 1992, AJ, 104, 340Makarova, L. 1999, A&AS, 139, 491Matthews, L. D., van Driel, W., & Gallagher, J. S. 1998, AJ, 116, 1169McGaugh, S. S. 1991, ApJ, 380, 140McMahon, P. M. 1993, Ph.D. Thesis, Columbia UniversityMirabel, I. F., & Wilson, A. S. 1984, ApJ, 277, 92Papaderos, P., Loose, H.-H., Thuan, T. X., & Fricke, K. J. 1996,

A&AS, 120, 207Patterson, R. J., & Thuan, T. X. 1996, ApJS, 107, 103Pilyugin, L. S. 2000, A&A, 362, 325Pilyugin, L. S. 2001, A&A, 369, 594Pilyugin, L. S., Molla, M., Ferrini, F., & Vılchez, J. M. 2002, A&A,

383, 14Rakos, K., Schombert, J., Maitzen, H. M., Prugovecki, S., & Odell, A.

2001, AJ, 121, 1974Richer, M. G., & McCall, M. L. 1995, ApJ, 445, 642Roberts, M. S., & Haynes, M. P. 1994, ARA&A, 32, 115Ronnback, J., & Bergvall, N. 1994, A&AS, 108, 193

Salpeter, E. E., & Dickey, J. M. 1985, ApJ, 292, 426Salzer, J. J., MacAlpine, G. M., & Boroson, T. A. 1989, ApJS, 70, 447Sandage, A., & Brucato, R. 1979, AJ, 84, 472Schommer, R. A., Sullivan III, W. T., & Bothun, G. D. 1981, AJ, 86,

943Shields, G. A., Skillman, E. D., & Kennicutt, R. C. 1991, ApJ, 371,

82Skillman, E. D., Bothun, G. D., Murray, M. A., & Warmels, R. H.

1987, A&A, 185, 61Skillman, E. D., Kennicutt, R. C., Shields, G. A., & Zaritsky, D. 1996,

ApJ, 462, 147Tarenghi, M., Garilli, B., & Maccagni, D. 1994, AJ, 107, 1629Thuan, T. X. 1985, ApJ, 299, 881Valluri, M., & Jog, C. J. 1990, ApJ, 357, 367van der Hulst, J. M. 1998, The Low Surface Brightness-Dwarf Galaxy

Connection, in Dwarf Galaxies and Cosmology, ed. T. X. Thuan,C. Balkowski, V. Cayatte, & J. Tran Thanh Van (Paris: EditionsFrontieres), 313

van Driel, W., & van Woerden, H. 1989, A&A, 225, 317van Driel, W., O’Neil, K., Cayatte, V., et al. 2003, A&A, 399, 433van Zee, L., Salzer, J. J., Haynes, M. P., O’Donoghue, A. A., &

Balonek, T. J. 1998, AJ, 116, 2805van Zee, L. 2000, AJ, 119, 2757van Zee, L. 2001, AJ, 121, 2003Vennik, J., Hopp, U., & Popescu, C.C. 2000, A&AS, 142, 399Vılchez, J. M. 1995, AJ, 110, 1090

J. Iglesias-Paramo et al.: A study of H -selected galaxies in the Hercules cluster, Online Material p 1

Online Material


ce-042

Fig. 5. Photometric properties of ce-042: Upper left: V-band image with superposed contours at 25.5, 24.5, 23.5, 22.5 and 21.5 mag arcsec−2.The coordinates are in J2000.0 Lower left: (V − I) colour map with V-band contours superposed Upper right: Surface brightness profiles.The fitted surface brightness distribution of the LSB component in B is depicted by the thick/grey curve. Open triangles show the emission inexcess of the LSB component Lower right: Colour profiles.


ce-048

Fig. 6. Same as Fig. 5 for ce-048.


ce-060



ce-061

Fig. 8. Same as Fig. 5 for ce-061. The ellipse indicates the segment of tidal tail included in the surface photometry analysis. The small crossesin the maps show a red pointlike sources (probably a foreground star) which was removed before computing the SBPs.


ce-143

Fig. 9. Same as Fig. 5 for ce-143. Due to strong overlapping with nearby objects, a 2D model was fitted and subtracted from the original imagesin order to be able to disentangle the light of the H -selected galaxy. Large uncertainties are expected in the SBP for µV > 25 mag arcsec−2.


ce-176



ce-200



sw-2 2 2

Fig. 12. Same as Fig. 5 for sw-222.


Appendix A: Comments on individual objects

Included are comments regarding the H properties, morpho-logical appearance and environment of the sample galaxies.

ne-112: Although there is nearby galaxy detected in H ata similar redshift, ne-176 at 5.′8 separation, no significant con-fusion is expected of the Arecibo H spectrum. The D97 VLAdata on ne-176 show VHI = 11 101 km s−1, W50 = 508 km s−1

and Iext = 0.81 Jykm s−1. In our spectrum its flux density con-tribution is expected to be at most a negligible ∼12% of theaverage flux density of ne-112.

ne-178: The uncertain optical redshift of 11 688 ±251 km s−1 listed in LEDA is based on two rather different mea-surements, 11 440 km s−1 (Hopp et al. 1995) and 11 935 km s−1

(Tarenghi et al. 1994). The VLA and Arecibo H values of,respectively, 11 556 and 11 591 km s−1 are rather closer to theformer optical value. The VLA data show a clear velocity gra-dient from the SW to the NE.

ne-204: The VLA data show a highly extended H distri-bution, but no mention is made of a velocity gradient in D97.

ne-208: The VLA detection is unresolved, with only13 pixels above threshold, i.e. an area of about half a beamsize.

ne-240: There are two nearby galaxies detected in H withredshifts similar to that of ne-240, ne-250, at 3.′5 separation,and ne-208, at 5.′8 separation, which are expected to causestrong confusion with the profile of ne-240. Except for the3.6 times higher Iext line flux of ne-250, the VLA profile param-eters of ne-240 and ne-250 are indistinguishable, seen the largevelocity resolution, while ne-208 has a 90 km s−1 lower veloc-ity and 0.6 times the VLA flux of ne-240. At most, we expectthe average flux density of nearby ne-250 and ne-208 in ourspectrum to be, respectively, ∼125% and ∼30% that of ne-240.

ne-250: Its H spectrum is not expected to be confusedsignificantly by that of nearby ne-240 (see above), whose lineemission is expected to be at most a negligible ∼11% of the av-erage flux density of ne-250. The VLA data show an extendedH distribution with a clear velocity gradient from the NW tothe SE.

ne-398: This reported VLA source was not confirmed bythe Arecibo data. The VLA detection has a peak line flux den-sity of 2.1 mJy in the integrated profile corresponding to theIH integrated line flux (see Sect. 3.2), while the estimated meanflux density of the profile corresponding to the I ext line fluxis 4.1 mJy. Such a line should have been easily detectable withthe 0.22 mJy noise level of our Arecibo data. However, insteadof a strong, 352 km s−1 wide profile centred on 10 602 km s−1,at Arecibo we only marginally detected a 364 km s−1 wide pro-file centred on 10 820 km s−1, with a 3.2σ peak after smooth-ing to a resolution of 19.5 km s−1. Although the reportedVLA source lies at the edge of the “ne” field, where the pri-mary beam attenuation factor is 4, its detection seemed real,according to D97. Our Arecibo data indicate it was spurious.

ce-042: The published VLA profile parameters are uncer-tain as the detection lies just at the edge of the band. TheVLA data show a velocity gradient from north to south. Ouroptical redshift of 11 840 ± 90 km s−1 agrees within the uncer-tainties with the 11 959 ± 60 km s−1 measured by Huchra et al.

(1995). The colour gradient was found to be γ+ ≈ 0 mag kpc−1

in both B−V and V−I for R∗ ≥ 2′′. Mean B−V and V−I coloursof 0.18 and 0.27 mag, respectively. The optical appearance ofce-042 is that of an Irregular galaxy. This object is almost iso-lated, with the elliptical galaxy PGC 057123 (BT = 16.81) asits closest bright neighbour, at a distance of about 4 ′ and witha radial velocity difference of 1000 km s−1.

ce-048: There are two nearby galaxies detected in H withredshifts similar to that of ce-048, which are expected to causeconfusion with its Arecibo profile: ce-060, at 2.′7 separation,and ce-095, at 6.′8 separation. There is only ±one VLA veloc-ity channel (44 km s−1) difference between their central H ve-locities and that of ce-048; the VLA line widths of ce-048 andce-060 are the same, 171 km s−1, while that of ne-95 is twiceas large. We expect the average flux density of nearby ce-060and ce-095 in our spectrum to be, at most, respectively, ∼66%and ∼50% of that of ce-048. The VLA data show a quite ex-tended distribution, with a velocity gradient from NW to SE.Its γ+ ≈= −0.05 mag kpc−1 in both colours, and its mean B−Vand V−I colours are 0.27 and 0.38 mag, respectively. It appearsas a faint blue galaxy, more compact than ce-042. There areseveral cluster galaxies near this object, the closest one beingPGC 057115 (BT = 16.84), a lenticular galaxy with 400 km s−1

difference in radial velocity, 1.′1 away.ce-060: Like in the case of ce-048 (see above), also for

this pointing position serious confusion is expected at Arecibobetween the H lines of ce-048, ce-060 and ce-095. We ex-pect the average flux density of nearby ce-048 and ce-095 inour spectrum to be, at most, respectively, ∼50% and ∼70% ofthat of ce-060. Its γ+ ≈= −0.04 mag kpc−1 in B − V (meancolour 0.27 mag), and it shows practically no V − I colourgradient, with a mean V − I colour of 1.04 mag. In its vicin-ity lies the bright Sab galaxy IC 1185 (BT = 14.89), at a dis-tance of 0.′7, with a velocity difference between both galaxiesof about 700 km s−1.

ce-061: An extended H distribution, measuring at least 2.′3in the E-W direction, surrounds the peculiar galaxy IC 1182(BT = 15.37, Vopt = 10 223 km s−1), formed to the East bythe VLA source ce-061 and to the West by the much weakersource ce-086. In the optical, IC 1182 shows a jet-like struc-ture towards the East, following the direction of the H distri-bution. The target galaxy ce-061 lies at the tip of the H tailextending eastwards from IC 1182, at about 1.′5 from the cen-tre of IC 1182. The CCD image of ce-061 (Fig. 8) showstwo distinct peaks, and the maximum in the H tail coincides,within the 25′′ beam size, with the easternmost optical peak.The SBP has been taking over the area covering both concen-trations seen in Fig. 8. A recent Hα line Fabry-Perot velocityfield (Duc & Amram 2003) shows that ce-061 is a gravitation-ally bound system inside the tidal tail of the IC 1182 mergersystem, confirming its interpretation as a tidal dwarf galaxy byBraine et al. (2001), who detected about 7 × 10 9 M� of H2

in a resolved distribution in IC 1182, but failed to detect ce-061 in the CO(1-0) or (2-1) lines, putting an upper limit to itsH2 mass of about 6 × 107 M�. We obtained Arecibo H spec-tra of ce-061 and ce-086, with pointing centres correspond-ing to the VLA H positions, which are 1.′8 (half the AreciboHPBW) apart. Though the two peaks that can be identified


with the VLA profiles of ce-061 and ce-086 are clearly presentin our spectrum, centered on ∼10 080 and 10 450 km s −1, re-spectively, the latter component is much stronger than in theVLA profiles: the average flux density in both components isabout 16 and 2 mJy at the VLA, using the I ext fluxes, whiletheir peak fluxes are ∼15 and 6 mJy in the Arecibo profile cen-tered on ce-061 and ∼9 and 6 mJy in the profile centered once-086. This may be due to extended emission in the distri-bution not included in the VLA profiles. The VLA data showa fairly smooth velocity gradient along the major axis of theH distribution and very little gas between the two peaks, at10 200–13 000 km s−1, in the area of IC 1182 itself. γ+ ≈=−0.013 mag kpc−1 in B−V (mean colour 0.22 mag); V−I colourgradient of 0.016 mag kpc−1, mean colour 0.56 mag. H detec-tions of IC 1882 were also reported by Bieging & Biermann(1983), Bird et al. (1993), Bothun et al. (1981, 1985), Mirabel& Wilson (1984), Salpeter & Dickey (1985) and Schommeret al. (1981). The average profile parameters from these refer-ences are VHI = 10 239±62 km s−1, W20 = 592±48 km s−1 andIHI = 3.6 ± 0.8 Jy km s−1. These values correspond well withour Arecibo parameters for ce-61 of 10 263 km s −1, 666 km s−1

and 3.4 Jy km s−1, respectively.ce-143: The CDD images show three, overlapping object.

The peak of the VLA H distribution corresponds to the objectin the middle, the weakest of the three. Before determining itsmorphological component, two-dimensional model light dis-tributions of its two neighbours were subtracted from the im-ages. Zero colour gradient, mean V − I colour of 0.48 mag.The optical counterpart of this galaxy seems to be a triple nu-cleated source, with the H cloud centred on the middle one.At a distance of 0.′7 of this system we find the Sa galaxyPGC 057055, with BT = 15.88, which shows a radial veloc-ity more than 700 km s−1 higher than ce-143.

ce-176: Although the VLA data show a galaxy in the vicin-ity with a similar redshift, ce-200, we do not expect our Areciboprofile to be confused with it: at 4.′4 distance, it has only a22 km s−1 higher central velocity than ce-176, but its meanVLA line flux is three times smaller. We expect the averageflux density of ce-200 in our spectrum to be, at most, a negligi-ble 7% of that of ce-176. The VLA data show a very extendedH distribution. The galaxy shows practically a zero gradientin both colour profiles, and mean B − V and V − I indicesof 0.42 mag and 0.66 mag, respectively. This galaxy appearslike a dwarf Irregular galaxy with a very extended H disc.Also, some knots of star formation are seen in the optical im-ages. The closest bright companion of this galaxy, at a distanceof 2.′5, is PGC 057020 (BT = 16.22), an SB0/Sa galaxy with aradial velocity difference of about 1000 km s−1 with respect toce-176.

ce-200: The Arecibo spectrum of this object shows threepeaks: one around 9780 km s−1 another around 9940 km s−1

and an about 300 km s−1wide feature around ∼10 200 km s−1.Though the centre peak is very close in velocity to ce-200,it is expected to be seriously confused with emission from

nearby ce-176 (see above). We expect the average flux den-sity of ce-176 in our spectrum to be, at most, equal to that ofce-200. No object can be found in the VLA data that mightcause the peak at 9780 km s−1, while the third peak appearsdue to ce-199, at 1.′6 distance, for which the VLA data showa VHI of 10 204 km s−1, a FWHM of 284 km s−1and a meanflux density of 1.5 mJy. γ+ ≈ 0 mag kpc−1 in B − V; meancolour 0.34 mag. The V − I profile shows a slight colourgradient of ≈0.03 mag kpc−1 and a mean value of 0.77 mag.No morphological peculiarities are apparent for this galaxy.Although it shows no signs of tidal interaction, two brightgalaxies appear close to it, showing also a similar radial ve-locity: NGC 6043A (V = 9879 km s−1 and BT = 14.10) andNGC 6045 (V = 9986 km s−1 and BT = 14.87), classified, re-spectively, as S0 and Sc, and located at distances of 1.′3 and 1.′6from ce-200.

sw-103: the VLA data show a velocity gradient from north-east to southwest. See also Sects. 3.2.4 and 4.

sw-194: The strong (713 mJy), extended continuum sourcein the nearby galaxy NGC 6034 made sensitive Arecibo lineobservations impossible and the VLA detection tentative. TheVLA profile is weak (0.75 mJy peak in the IH profile) and thereare only 15 pixels above the detection threshold. The “line”could be a figment of imperfect bandpass calibration and con-tinuum subtraction, as the continuum is sufficiently strong toincrease the noise in the spectral baselines in its vicinity (D97).See also Sects. 3.2.4 and 4.

sw-222: The VLA data show an extended H distribution,with a plume extending about 1 ′ towards the SE from the cen-tre, and a velocity gradient along the main body from west toeast. γ+ ≈ 0 mag kpc−1 in V − I; mean colour 0.68 mag. Thisgalaxy shows an elongated shape with an H plume extendingtowards the South. The closest bright galaxy, which is located4 arcmin away, is the lenticular PGC 056824 (BT = 14.00),whose velocity difference with sw-222 is about 300 km s−1.

47-138 No sensitive Arecibo observations could be ob-tained due to the presence of the 1.2 Jy radio source 4C+15.53,whose centre is just 40′′ offset from the galaxy. It could be de-tected at the VLA as the continuum source has a diameter ofabout 100′′ and the VLA peak flux density, 50 mJy/beam, islow enough not to disturb the H detection (1.1 mJy peak inthe IH profile). The VLA data show a velocity gradient fromnorth to south.

47-154 This weak VLA source was not detected atArecibo. The VLA detection has a peak line flux of 0.7 mJyin the integrated profile corresponding to the I H integrated lineflux (see Sect. 3.1), while the estimated mean flux density ofthe profile corresponding to the Iext line flux is 0.37 mJy. Itis classified as a somewhat tentative detection in D97. Withan 0.32 mJy rms noise level in our Arecibo data, this sourceappears too weak for detection or confirmation.

47-211 The VLA data show a velocity gradient from NEto SW, along the major axis of the galaxy.

Documents

A study of H I-selected galaxies in the Hercules cluster